Tight Junction-Mediated Morphologic and Adherent Diversities Act on Hematopoietic Fate Induced by OCT4 in human hair follicle mesenchymal stem cells

Background: Human hair follicle mesenchymal stem cells (hHFMSCs) isolated from hair follicles possess multilineage differentiation potential. OCT4 is a gene critically associated with pluripotency properties. The cell morphology and adhesion of hHFMSCs signicantly changed after transduction of OCT4 and two subpopulations emerged, including adherent cells and oating cell. Floating cells cultured in hematopoietic induction medium and stimulated with erythropoetic growth factors could transdifferentiate into mature erythrocytes, whereas adherent cells formed negligible hematopoietic colonies. The aim of this study was to reveal the role of cell morphology and adhesion on erythropoiesis induced by OCT4 in hHFMSCs and to characterize the molecular mechanisms involved. Methods: Floating cell were separated from adherent cell by centrifugation of the upper medium during cell culture. Cell size was observed through ow cytometry and cell adhesion was tested by disassociation and adhesion assays. RNA sequencing was performed to detect genome-wide transcriptomes and identify differentially expressed genes. GO enrichment analysis and KEGG pathway analysis were performed to analysis the functions and pathways enriched by differentially expressed genes. The expression of tight junction core members was veried by qPCR and Western blot. Results: The overexpression of OCT4 inuenced the morphology and adhesion of hHFMSCs. Transcripts in oating cells and adherent cells are quite different. Data analysis showed that upregulated genes in oating cells were mainly related to the pluripotency, germ layer development (including hematopoiesis lineage development), and downregulated genes were mainly related to cell adhesion, cell junctions and the cytoskeleton. Most molecules of the tight junction (TJ) pathway were downregulated and molecular homeostasis of the TJ was disturbed, as CLDNs were disrupted, and JAMs and TJPs were upregulated. The dynamic of cell adhesion-related gene E-cadherin and cytoskeleton-related gene ACTN2 might cause different morphology and adhesion. Finally, a regulatory network centered to OCT4 was constructed, which elucidated the TJ pathway critically bridges pluripotency and hematopoiesis in a TJP1-dependent way. Conclusions: Regulations of cell morphology and adhesion via the TJ pathway conducted by OCT4 might modulate hematopoiesis in hHFMSCs, thus developing potential mechanism of erythropoiesis in vitro. hHFMSCs oating displayed hematopoietic differentiation suggest the possible effects of changes in cell adhesion, junctions, and on pluripotency and hematopoietic differentiation in OCT4-reprogrammed hHFMSCs. RIN:RNA integrity number; FC:fold change; DEGs:differentially expressed genes; KEGG:Kyoto Encyclopedia of Genes and Genomes; GO:Gene Ontology.


Introduction
Erythropoiesis is a stepwise process through which red blood cells (RBCs, erythrocytes) are generated from hematopoietic stem and progenitor cells (HSPCs) and is controlled by multiple elements. Inducing erythrocyte production in vitro provides a model system for exploring the mechanisms of erythropoiesis. Previously, a population of small round oating cells with subtle expression of hematopoietic stem cell (HSC) marker CD45, gradually emerged from OCT4(POU5F1)-reprogrammed human hair follicle mesenchymal stem cells (hHFMSCs OCT4 ), and could transdifferentiate into mature enucleated RBCs when stimulated with a combination of hematopoietic cytokines (1). This prompted us to consider an association between this particular cell morphology and adhesion with possible erythropoiesis mechanisms, that is, low adhesion and round-like cell morphology conferring higher hematopoietic capacity to hHFMSCs OCT4 when treated with cytokines, thus promoting transduction of cellular signals and subsequently initiating the process of erythropoiesis.
It is well known that blood cells grow in suspension, and the process of erythropoiesis is accompanied by great changes in cell morphology. The cell size gradually increases as hematopoietic progenitor cells (HPCs) differentiate into precursors (2), and then decreases during erythroblast maturation accompanied by cytoskeleton remodeling and loss of cytoplasmic-nuclear connections (2,3). Moreover, some studies have revealed that the self-renewal and differentiation of HSCs are affected by cell morphology and adhesion. Rho kinases controlling the cytoskeleton are required for the biological functions of HSCs, and are particularly signi cant for enucleation during erythropoiesis (4)(5)(6). Cell morphology and adhesion affect the polarity and proliferation of HSPCs (7)(8)(9). Platelet factor 4 binds to HPCs, strengthens the adhesion of HPCs to the extracellular matrix, and ultimately modulates hematopoiesis (10). However, whether the characteristics of a speci c cell morphology and low adhesion are crucial factors in erythropoiesis and how they facilitate RBC development are still obscured.
Cell junction molecules are a type of cell adhesion molecules, and cell adhesion and cell junction systems dynamically and mutually interact (11). Furthermore, tight junctions (TJs) play a role in recruiting various cytoskeleton and signaling molecules on their cytoplasmic surface, and linking extracellular proteins with intracellular signaling pathways to the cytoskeleton (12). TJ members could affect the biological functions of HSCs. For example, intimate intercellular contact mediated by JAMs is required for e cient transduction of Notch signaling in HSCs, and de ciency of JAM1 results in impaired HSC speci cation (13,14). JAM2 regulates the maintenance of HSCs through heterotypic interactions with JAM3 (15,16), and depletion of JAM3 leads to a sharp decrease in the frequency of myeloid progenitors in the bone marrow (17). It is worth noting that the master hematopoietic regulator RUNX1 can respectively bind to TJP1, OCLN and CLDN5 via the "TGGGGT" DNA sequence in the promoter region (18). Overall, the overexpression of OCT4, the most important pluripotent transcription factor (TF), alters the morphology and adhesion of hHFMSCs, making the cells more prone to cytokine stimulation and differentiate towards the hematopoietic lineage.
In this study, we utilized next-generation sequencing combined with bioinformatics analysis, qPCR, and Western blotting, to investigate the interactions between cell adhesion, cytoskeleton, hematopoiesis and pluripotency modulated by OCT4 in hHFMSCs. Ideally, elucidating the molecular mechanisms of OCT4reprogrammed hHFMSCs differentiation into erythrocytes will not only help to understand the mechanism of RBC production, but also provide an experimental basis for hematopoiesis in vitro for future clinical applications.

Materials And Methods
Cell Culture hHFMSCs were isolated from root tissue of hair follicles cultured in 96-well plates (19), hHFMSCs OCT4 were obtained by lentivirus transduction of OCT4 into hHFMSCs, and the cells were seeded onto Matrigelcoated culture plates. Floating cells ( oating hHFMSCs OCT4 ) were sorted from the upper medium of adherent cells (adherent hHFMSCs OCT4 ) by centrifugation during culture. All cells were maintained in H-DMEM/F12 medium (Gibco) containing 10% fetal bovine serum (Gibco), 100 U/mL penicillinstreptomycin (HyClone) and 10 ng/mL broblast growth factor-basic (Acro Biosystems) at 37 °C and 5% CO 2 in a cell culture incubator.

Flow Cytometry Analysis
A total of 10 7 cells/mL were prepared for ow cytometry analysis. A 200µL aliquot was added to 96-well plates per well, and then the light-scattering properties of the cells were measured by ow cytometry. In the ow cytometry assay, the value of forward scatter (FSC) is proportional to the size of the cells, so the FSC can be used to compare the relative size of the cells. FSC distribution histogram plots were generated by a computer with raw ow cytometry data.

Dissociation Assay
Cell dissociation assays were performed as previously described (20). Cells were treated with dispase II (2.4 U/mL, Sigma Aldrich) to disassociate cells from culture plates with minimal destruction of intercellular junctions. The numbers of single and total cells were counted with a hemocytometer. The percentage of individual cells to the total number of cells was inversely proportional to cell-to-cell adhesion. Each sample was tested three times.

Cell Adhesion Assay
Cell-extracellular matrix adhesion was assayed according to Codogno et al with minor modi cations (21). The same number of cells was seeded onto Matrigel-coated 24-well plates and incubated for 2 hours. The plates were washed three times to remove unattached cells and the cell number was counted using a hemocytometer. The percentage of remaining cells to the total number of cells was proportional to cellextracellular matrix adhesion.

RNA-seq and DEG Analysis
Three independent biological replicates of the same sample were sent to Shanghai Oe-biotechnology (Shanghai, China) for RNA-seq analysis. Total RNA was extracted using the mirVana miRNA Isolation Kit and RNA integrity number (RIN) was evaluated by an Agilent 2100 Bioanalyzer. Samples with RIN greater than or equal to 7 were sequenced by an Illumina HiSeq X Ten sequencer (Illumina). The ltered clean reads were mapped to the reference genome database (accession number: GCF_000001405.38) by hisat2, v2.2.1.0 (http://ccb.jhu.edu/software/hisat2/index.shtml). Each transcript was normalized by FPKM to eliminate the in uence of gene length and sequencing depth. The counts of each sample were mapped to the annotated genome after standardization and normalization. Finally, fold change (FC) and difference signi cance were used to screen the differentially expressed genes (DEGs). DEGs with FC value greater than 2 or lower than − 2, and a P-value lower than 0.05 were considered signi cant.
GO term and KEGG pathway enrichment analyses were carried out using the tool for Function Annotation in the DAVID (https://david.ncifcrf.gov/).The KEGG pathway maps were obtained from the KEGG database (http://www.kegg.jp/). Signi cant genes were visualized by the STRING database (http://stringdb.org/), and a network was constructed using Cytoscape software (https://cytoscape.org/).

Expression Validation using qPCR
Total RNA was extracted from 5 × 10 6 cells treated with 1 mL TRIzol (Sparkjade, Shandong, China), and the purity and concentration were determined by a NanoDrop 2000 (Thermo Fisher Scienti c). cDNA was synthesized with the PrimeScript RT reagent Kit (+ gDNA Eraser) and then subjected to qPCR using TB Green® Premix Ex Taq™ II (Takara). The gene mRNA levels were determined using 50 ng of cDNA on an Applied Biosystems 7300. All template ampli cations were conducted in triplicate with a three-step PCR process, which included one cycle of 95 °C for 30 seconds, 40 cycles of 95 °C for ve seconds and 60 °C for 31 seconds and one nal cycle of 95 °C for 15 seconds, 60 °C for one minute and 95 °C for 15 seconds. Using GAPDH expression as a normalization control, the relative expression was calculated as 2 −∆∆Ct . The primer sequences are provided in Table 1.

Western Blot Analysis
Total proteins were isolated with Radio Immunoprecipitation Assay buffer (Solarbio) supplemented with 1% Phenylmethylsulfonyl uoride (Solarbio), followed by centrifugation to remove cell debris. Then protein extracts were subjected to protein estimation using a BCA protein assay kit (Solarbio). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried with equal amounts of proteins, and then proteins were transferred to a polyvinylidene uoride membrane (Millipore). Upon blocking with 5% nonfat milk, the membrane was incubated with primary antibody overnight at 4 °C. After washing three times with TBST, the membrane was incubated with HRP-conjugated secondary antibody at room temperature for 2 h. Finally, the blots were developed with Immobilon Western Chemilum HRP Substrate (Millipore). Differential protein expression was calculated as a ratio normalized to β-actin protein expression.

Statistical Analysis
Data comparing two groups were statistically analyzed by Student's t-test. GraphPad Prism 8.0.2 was used to generate histograms. P-values inferior to 0.05 were considered signi cant.

Overexpression of OCT4 Causes Changes in Morphology and Adhesion in hHFMSCs
The morphology of OCT4-reprogrammed hHFMSCs greatly changed in the process of erythropoietic differentiation, whereby the spindle-shaped cells became polygonal, and a population of small oating round or quasi-round cells emerged from the adherent polygonal cells (Fig. 1a). We then determined the relative size of these cells by ow cytometry. The data showed that the size of adherent hHFMSCs OCT4 was notably smaller than that of hHFMSCs, while oating hHFMSCs OCT4 were smaller than the adherent cells ( Fig. 1b), which was consistent with what we observed under an optical microscope. At the same time, cell adhesion changed as a portion of the cells gradually became suspended in the medium, so dissociation and adhesion assays were carried out to detect cell adhesion. The percentage of single cells was higher in adherent hHFMSCs OCT4 (47.5%) than in hHFMSCs (9.4%), and the percentage was higher in oating hHFMSCs OCT4 (85%) than in adherent hHFMSCs OCT4 (Fig. 1c). In the adhesion assay, the percentage of remaining cells was lower in adherent cells (18.4%) than in hHFMSCs (70.9%), while it was lower in oating hHFMSCs OCT4 (2.3%) than in adherent hHFMSCs OCT4 (Fig. 1d). The above results validated that the morphology of OCT4-reprogrammed hHFMSCs changed and adhesion decreased. It is worth noting that it was the population of oating hHFMSCs OCT4 with low-adhesion prone to transdifferentiate towards erythroid lineage after stepwise stimulation by cocktails of hematopoietic cytokines. Accordingly, cell morphology and adhesion might be the negative factors affecting erythropoiesis.

Transcripts in Cells with Diverse Morphology and Adhesion are Quite Different
To investigate the role of cell morphology and adhesion during erythrocyte differentiation from hHFMSCs OCT4 , RNA-seq was performed and the DEGs were sorted. First, to compare the similarity of samples within a group and the diversity of each group, the principle components were analyzed. As shown in the three-dimensional distribution (Fig. 2a), the closer distances of cells within each group implied good repetitiveness, and the distances between every two groups were signi cantly greater, especially when adherent hHFMSCs OCT4 and oating hHFMSCs OCT4 were compared with hHFMSCs, respectively. Based on these data combined with the correlation coe cient and clustering of correlation provided in Additional le 1, OCT4 induced hHFMSCs to derive novo cell populations, remarkably, cells with different morphology and adhesion might possess distinct gene transcripts.
Next, DEGs were sorted among the three groups. When we compared adherent hHFMSCs OCT4 with hHFMSCs, 2401 upregulated genes and 1882 downregulated genes were identi ed (Fig. 2b), and 3107 upregulated genes and 2999 downregulated genes were determined in oating hHFMSCs OCT4 compared to hHFMSCs (Fig. 2c), indicating that OCT4 conferred considerable changes of transcriptome in the whole genome to hHFMSCs. Importantly, 833 upregulated genes and 1107 downregulated genes were also identi ed when oating hHFMSCs OCT4 were compared with adherent hHFMSCs OCT4 (Fig. 2d). Although the number is smaller, it would de nitely play a considerable role in oating cells. Venn diagram analysis revealed a total of 612 and 388 group-speci c DEGs for adherent hHFMSCs OCT4 vs. hHFMSCs and oating hHFMSCs OCT4 vs. adherent hHFMSCs OCT4 , respectively (Fig. 2e). In particular, 1785 groupspeci c DEGs were identi ed in oating hHFMSCs OCT4 vs. hHFMSCs, which was a much larger number than that in the other two comparison groups. This considerable number of DEGs probably bring about signi cant changes in biological function to OCT4-reprogrammed hHFMSCs. Especially, the groupspeci c DEGs may yield tremendous changes to oating hHFMSCs OCT4 when compared with hHFMSCs. In the DEGs analysis, it is likely that the common and group-speci c DEGs collectively affected the morphological characteristics and subsequent transdifferentiation.

Floating Cells Lose Part of Pluripotency and Gain Hematopoietic Differentiation Potential
Transduction of the key pluripotent TF OCT4 would de nitely in uence the pluripotency of hHFMSCs. Consequently, we focused our analysis on the expression of related genes in cells with different morphology and adhesion. hHFMSCs expressed negligible levels of OCT4, LEFTY2, SOX18, POU3F2 and SEMA4D (Additional le 2: Table S1), and both adherent hHFMSCs OCT4 and oating hHFMSCs OCT4 expressed higher levels of pluripotent genes, including LEFTY2, KLF4, MYC, POUs, SEMAs and SOXs, than hHFMSCs (Fig. 3a). Some of the pluripotent genes, such as LEFTY2, SOX4, and SEMA6C, however, were downregulated in oating hHFMSCs OCT4 compared with adherent hHFMSCs OCT4 (Fig. 3a), which was validated by KEGG enrichment analysis as downregulated genes were enriched in the term signaling pathways regulating pluripotency of stem cells in oating hHFMSCs OCT4 vs. adherent hHFMSCs OCT4 (Fig. 3b). These results suggested that OCT4-reprogramed hHFMSCs acquired pluripotency but lost some of it after adherent cells transformed into the oating subset. The DEGs were then clustered according to their GO terms using DAVID, and the top 10 GO terms related to differentiation and development enriched with upregulated genes and downregulated genes were separately analyzed. Upregulated DEGs were involved in the terms of germ layer differentiation in adherent hHFMSCs OCT4 and oating hHFMSCs OCT4 when compared with hHFMSCs (Table 2 and Additional le 2: Table S2, S3), and the upregulated genes in oating cells were specially enriched in the terms T-helper 1 cell differentiation and regulation of erythrocyte differentiation (Table 2), implying a potential for erythropoietic differentiation.

Downregulation of the Tight Junction Pathway in Floating hHFMSCs OCT4
To further explore the role of cell morphology and adhesion during erythropoiesis in hHFMSCs OCT4 , the top 10 GO terms, covering biological process, molecule function and cellular component, are displayed in Fig. 4a. Downregulated genes were obviously enriched in relevant terms,such as cell adhesion, focal adhesion, cytoskeleton and cell-cell junction in adherent hHFMSCs OCT4 compared to hHFMSCs. In addition, downregulated genes were signi cantly enriched in the term cell adhesion in oating hHFMSCs OCT4 relative to adherent hHFMSCs OCT4 , suggesting the sharp decrease in adhesion of oating hHFMSCs OCT4 . Besides, KEGG analysis revealed that downregulated genes were enriched in cell signaling pathways including regulation of actin cytoskeleton, cell adhesion molecules and focal adhesion in adherent hHFMSCs OCT4 vs. hHFMSCs, as well as pathways of regulation of actin cytoskeleton, gap junction, adherens junction and focal adhesion in oating hHFMSCs OCT4 vs. adherent hHFMSCs OCT4 (Fig. 3b). These results veri ed the changes in cell morphology and adhesion of hHFMSCs OCT4 , which were consistent with our observations. Therefore, the morphology-and adhesion-related genes aroused corresponding alterations in OCT4-reprogrammed hHFMSCs and facilitated the switch between adherent and oating subpopulations.
Tight junctions, generally known for their fence function controlling cellular matter diffusion, can also modulate cell adhesion and the cytoskeleton (12,22). The TJ pathway was found to be downregulated by KEGG analysis in these three comparison groups, and the TJ pathway was annotated through the KEGG database and the DEGs were annotated (Fig. 4b). There were 12 upregulated genes and 20 downregulated genes in the TJ pathway. The results clearly showed that TJ genes were dynamically expressed, and several programs, such as cell proliferation, adhesion, cytoskeleton, cell polarity, paracellular permeability and most importantly cell differentiation, were involved. Furthermore, uctuation of one member in the TJ pathway would inevitably affect other member molecules and thereby have an impact on the biological functions of hHFMSCs OCT4 and initiate the switch between the two states characterized by different morphology and adhesion.

Gene Expression Validation by qPCR and Western blot
Cell junction molecules, including TJ members are involved in cell adhesion and could directly affect cell adhesion (11). Therefore, we performed qPCR to detect the mRNA expression levels of selected genes associated with TJs, adhesion or cytoskeleton (Fig. 5a). As expected, the expression level of the TJ member gene CLDN11 was signi cantly decreased in adherent hHFMSCs OCT4 and oating hHFMSCs OCT4 relative to hHFMSCs, but both CLDN6 and CLDN7 were increased. Especially, CLDN5 was downregulated in adherent hHFMSCs OCT4 and then upregulated in oating hHFMSCs OCT4 . The expression levels of JAM1 and JAM3, which play an important role in the commitment of lineage speci cation and cellular signaling transduction in HSCs, were respectively decreased and increased in oating hHFMSCs OCT4 vs. hHFMSCs.
Moreover, the expression levels of TJP1, TJP2 and TJP3, core members associated with the cytoskeleton and intracellular signaling transduction, were remarkably upregulated 5.4-fold, 59.4-fold and 7.6-fold in oating hHFMSCs OCT4 . These results indicated disrupted molecular homeostasis of the TJ pathway upon OCT4 transduction. We also found that the cytoskeleton gene ACTN2 increased more than 20-fold in oating cells, and the expression of E-cadherin increased in adherent hHFMSCs OCT4 and then decreased in oating hHFMSCs OCT4 , implying that the changes in adhesion of these cells might be related to calcium signals.
Changes in cell morphology and adhesion in the process of erythropoiesis could in uence the biological functions of hematopoietic cells, including proliferation, self-renewal, differentiation and etc. (8), we also detected the expression levels of the terminal erythroid differentiation-related gene ROCK1 and the essential hematopoietic development gene RUNX1. Expression of these two genes were found decreased in oating cells, indicating that the state of hematopoietic program was not yet triggered, although OCT4 conferred pluripotency in this group of cells to some degree as other pluripotency genes MYC, KLF4, etc. were upregulated as shown in previous sequencing data.
Next,Western Blotting was carried out to validate the protein expression levels of TJP1, JAM1, CLDN5, CLDN11 and RUNX1. As shown in Fig. 5b, the expression tendency of these proteins was consistent with the mRNA levels. The above results corresponded to our sequencing data, except for TJP1, which was remarkably increased in adherent cells and oating cells compared to hHFMSCs but was not identi ed by sorting DEGs.

A Putative Regulatory Network in Floating hHFMSCs OCT4
Although pluripotency-, cell morphology-and cell adhesion-related genes have been identi ed, the internal correlations between these factors are still unknown. Therefore, we visualized the signi cant DEGs in oating hHFMSCs OCT4 vs. hHFMSCs and constructed a network (Fig. 6a). This network suggested that pluripotency-related genes in hHFMSCs, such as OCT4, SOX2, c-MYC (MYC) and KLF4, were regulated by genes related to cell adhesion, cell junction, and cytoskeleton, including TJP1, TJP2, FN1 ( bronectin 1), CTNNB1 (β-catenin), CDH1 (E-cadherin), ACTB (β-actin) and ACTG1 (γ-actin 1). In addition, there were interactions within the TJ pathway, as well as interplay between the TJ pathway, cell adhesion-and cytoskeleton-related molecules, such as PTK2 (FAK), CDH1, CTNNB1, ACTB, and ACTG1. Hematopoietic genes, such as, CD44, CD117 (KIT), RUNX1 and ROCK1 could similarly bind to or interact with cell adhesion and junction molecules. These results strongly imply that dynamic expression of genes related to adhesion, cytoskeleton and junctions results in the low-adhesion and different morphology of hHFMSCs OCT4 , whereas oating cells lost some of the pluripotency and became more prone to differentiation into blood cells. Furthermore, TJP1 uniquely links pluripotency, hematopoiesis and cytoskeleton with the TJ pathway through KLF4, RUNX1, ACTB and ACTG1. Fluctuation of one member in TJ pathway would inevitably affect other member molecules, thereby impacting the biological functions of hHFMSCs OCT4 and initiating the transform between the two states of different morphology and adhesion characteristics.
In summary, transduction of OCT4 brought about great differences in morphology and adhesion to hHFMSCs, whereby two subsets of cells appeared and gained pluripotency, with oating cells losing some of their pluripotency. The dynamically expressed TJ pathway before erythropoietic inducement might act as a pivotal point of changes in cell morphology and adhesion, resulting in damaged pluripotency in oating cells and probable constructed hematopoietic capacity in a TJP1-dependent way.

Discussion
It was reported that oating hHFMSCs OCT4 treated with hematopoietic cytokines could transdifferentiate into RBCs (1). Here, we investigated the correlation between cell morphology, adhesion, pluripotency and hematopoiesis to elucidate the mechanisms of erythropoiesis in OCT4-reprogrammed hHFMSCs. The results veri ed the alterations in morphology and adhesion in hHFMSCs OCT4 , and the corresponding changes in gene expression were detected by RNA-seq, qPCR and Western blot. In-depth analysis of sequencing data showed that the expression of genes related to pluripotency changed in hHFMSCs OCT4 with diverse adhesion and morphology, and oating cells displayed hematopoietic differentiation potential. All these results suggest the possible effects of changes in cell adhesion, junctions, and cytoskeleton on pluripotency and hematopoietic differentiation in OCT4-reprogrammed hHFMSCs.
Pluripotency-related DEGs dynamically expressed in two subpopulations of hHFMSCs OCT4 , and pluripotency was presumably reduced in oating cells compared with adherent cells. In humans, direct conversions of mature somatic cells to another type of cells using the single factor OCT4, a key pluripotency TF, have been widely developed (23,24). Although not shown to be involved in physiological hematopoiesis, OCT4 is capable of promoting the expression of essential hematopoietic regulators in supportive culture conditions (23). hHFMSCs are easily available and nonimmunogenic, making them a potential alternative stem cell source for patient-speci c applications. Transduction of the individual factor OCT4 allows hHFMSCs to transdifferentiate into RBCs under multiple hematopoietic induction conditions, providing an optional way to generate RBCs in vitro for transfusion (1). Evolutionary conservation analysis was performed to identify a relative subset of targets of OCT4 and other POU proteins that link the regulation of cell-cell adhesion to differentiation (25). Whether OCT4 regulates pluripotency through adhesion or regulates both in parallel is di cult to determine, however, analysis of the conserved OCT4 network indicates that the regulation of differentiation and adhesion is inseparable (25). This may partially explain the relatively decreased pluripotency in oating hHFMSCs OCT4 .
Attenuated pluripotency prompts cells to be more likely to differentiate into a certain cell type. Blood cells are well known to grow in suspension and erythropoietic precursors show a progressive reduction in cell and nuclear size during erythropoiesis (26), extremely similar to the sequencing results and our observations upon oating hHFMSCs OCT4 . A number of molecules associated with cell adhesion and the cytoskeleton in hematopoietic cells are proved essential for erythropoiesis under homeostasis and stress (27)(28)(29)(30)(31). Embryonic hematopoiesis involves activation of a hematopoietic transcriptional program, followed by major morphological changes and breakage of the tight junctions (32). Tight junction members, TJPs, CLDNs and JAMs control self-renewal, proliferation and recovery from stress in hematopoietic cells, including ESCs, HSCs and pluripotent stem cells (17,18,(33)(34)(35)(36)(37). Our results of sequencing analysis, qPCR and Western blot clearly demonstrate dynamically expressed TJ members in hHFMSCs OCT4 with divergent morphology and cell adhesion. Mutual conversion between adherent hHFMSCs OCT4 and oating hHFMSCs OCT4 seems to account for the dynamic TJ pathway, causing disparate morphology and adhesion, and ultimately regulating the expression of genes related to pluripotency and hematopoiesis (Fig. 6b). However, it is still unclear whether these morphological alterations initiate the process of erythropoiesis, or are the inevitable phenomenon accompanying erythroid hematopoiesis since erythropoiesis requires membrane biogenesis, establishment of cell polarity and cytoskeleton assembly during differentiation and enucleation (38)(39)(40). More research should be conducted to address this problem.
In mammals, RUNX1 is identi ed as a core regulator of hematopoiesis that is essential for the initiation of the hematopoietic program in embryonic hematopoiesis (41)(42)(43), and there is bidirectional negative regulation between TJP1 and RUNX1 (18,44). Therefore, TJP1 suppressed the expression of RUNX1 in oating cells, implying insigni cant hematopoietic differentiation tendency, while the expression of TJPs (TJP1, TJP2 and TJP3) was upregulated in oating cells. Consequently, the morphological changes potentially confer a certain degree of pluripotency to oating hHFMSCs OCT4 , accompanied by probable enhanced sensitivity to hematopoietic cytokines and thus trigger the hematopoietic program.
Herein we present a hypothetical model (Fig. 6b), in which the TJ pathway might regulate hematopoiesis in OCT4-reprogrammed hHFMSCs by transforming cell morphology from adherent polygonal or fusiform to oating small and round. Although further experiments should be followed up, such as knock-out of TJP1, investigations of pluripotency and hematopoietic capacity, study of morphological changes effect on intercellular signal transduction and detection of sensitivity to hematopoietic cytokines.

Conclusions
We have developed potential mechanisms of erythropoiesis in OCT4-reprogrammed hHFMSCs, where changes in cell morphology and adhesion were involved through the TJ pathway. This study characterized possible hematopoietic molecular mechanisms in hHFMSCs OCT4 in vitro, providing comprehensive insight into the potential role of the TJ pathway during erythropoiesis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.